A Convolutional Neural Network for Multiple Particle Identification in the MicroBooNE Liquid Argon Time Projection Chamber MicroBooNE Collaboration June 2020 MICROBOONE-NOTE-1080-PUB Abstract We present the multiple particle identification (MPID) network, a convolutional neural network (CNN) for multiple object classification, developed by MicroBooNE. MPID pro- vides the probabilities of e−, γ, µ−, π±, and protons in a single liquid argon time projection chamber (LArTPC) readout. The network extends the single particle identification network previously developed by MicroBooNE [1]. MPID takes as input an image either cropped around a reconstructed interaction vertex or containing only activity connected to a recon- structed vertex, therefore relieving the tool from inefficiencies in vertex finding and particle clustering. The network serves as an important component in MicroBooNE's deep learn- ing based νe search analysis. In this note, we present the network's design, training, and performance on simulation and data from the MicroBooNE detector. Contents 1 Introduction 2 2 Multiple particle convolutional neural network 4 2.1 Network design . .4 2.2 Training and Test Samples . .6 2.3 Network Training . .7 2.4 MPID Occlusion Analysis . .7 3 Performance on Simulation 9 3.1 1µ-1p Simulated Sample . 10 3.2 1e-1p Simulated Sample . 11 4 Comparison of Data/Simulation Performance 15 4.1 1µ-1p Enriched Data . 15 0 4.2 νµCCπ Enriched Data . 19 1 5 Use of MPID In A Low Energy Excess Measurement 20 0 5.1 Simulated Intrinsic νe vs. νµCCQE and νµπ ..................... 20 5.2 Simulated Intrinsic νe vs. Cosmic Event . 23 6 Conclusion 24 1 Introduction A series of liquid argon time projection chamber (LArTPC) detectors have been or are being deployed at Fermilab as part of the Short-Baseline Neutrino (SBN) program [2] along the Booster Neutrino Beamline (BNB [3]) and as part of the long-baseline program of the Deep Underground Neutrino Experiment (DUNE) [4]. The MicroBooNE experiment [5], part of the Fermilab SBN program, has been operating since 2015, collecting data accumulated during beam-on and beam-off time periods. MicroBooNE operates a 170 ton (87 ton active) LArTPC placed 470 m from the BNB target at Fermilab. The LArTPC is 10.3 m long, 2.5 m wide and 2.4 m high. The detector has three readout wire planes with 2400 readout wires on the two induction planes and 3456 readout wires on the collection plane [6]. Wires are installed with two induction planes oriented at ±60◦ with respect to the vertical collection plane at a wire pitch of 3 mm. An array of 32 PMTs are installed behind the collection plane to detect the scintillation light from argon ionization caused by charged final state particles from neutrino interactions [7]. The TPC readout time window is 4.8 ms and is digitized into 9600 readout time ticks. Charged particles in liquid argon produce ionization electrons, which drift to the readout wire planes in an electric field of 273 V/cm. It requires 2.3 ms for an ionization electron to drift across the full width of the detector. The MicroBooNE LArTPC continuously records charge drifted and its arrival time on each wire. A software trigger, based on PMT signals, records an event triggered by the BNB beam spill if the interaction light detected by the PMT array is above a set threshold. Each event consists of data collected from 1.6 ms before the trigger and 3.2 ms after the trigger. Therefore, each event has three sets of TPC data for each wire on all three planes. The two induction planes have resolutions of 2400 wires × 6048 readout ticks, while the collection plane has a resolution of 3456 wires ×6048 readout ticks. Wire and time data can be converted into an image format (charge on each wire versus drift time) using the software toolkits LArSoft [8] and LArCV [9] while maintaining high resolution in wire, time and charge amplitude space. These information-rich LArTPC images are suitable for applying deep learning tools. In consideration of computing resources, images for deep learning tools are compressed along the time tick axis by a factor of six. Pixel values are merged by a simple sum. Images become 2400 wires × 1008 ticks and 3456 wires × 1008 ticks for the induction and collection planes, respectively. This corresponds to an effective position resolution of 3.3 mm [10] and 3 mm [6] along the time tick and wire number directions, respectively. Convolutional neural networks (CNN), deep learning networks commonly applied to image pro- cessing applications, are currently used across neutrino and high energy physics experiments [11]. For accelerator neutrino experiments, NOvA has applied a CNN as a neutrino event classifier [12] 0 in its νµ ! νe oscillation measurement [13, 14] and its neutral-current (NC) coherent π produc- tion measurement [15]. NOvA has also demonstrated a context-enriched particle identification network [16]. MINERvA has developed CNN tools to determine neutrino interaction vertices and study possible biases due to models used in the large simulated training sample [17]. The NEXT experiment has also used a CNN classifier to perform particle content studies at candidate 2 neutrinoless double beta decay vertices [18]. A variety of deep learning techniques have been used in neutrino LArTPC experiments. In MicroBooNE, a CNN for assigning probabilities of particle identities for single particles in the MicroBooNE LArTPC has been demonstrated on simulated data in Ref. [1]. A semantic segmen- tation network for LArTPC data [19] has been used for π0 event reconstruction [20], vertex finding, and track reconstruction [21]. The DUNE experiment has recently presented an updated long- baseline neutrino oscillation sensitivity study incorporating a CNN for neutrino event selection and background rejection [22]. In this article, we present our study in developing and applying a multiple particle identifi- cation (MPID) network with the task of multiple binary logistic regression problem solving in MicroBooNE. It is the first demonstration of the performance of a CNN on LArTPC data in- cluding systematic uncertainties, and the first particle identification network applied to LArTPC datasets. The MPID network extends the functionality of MicroBooNE's previously-described single PID CNN network [1]. It does not require pre-processing of image data to identify and filter selected pixels in an image assumed to be produced by a specific particle. The network provides simultaneous prediction scores for particle existence probabilities in the same image among five different particle species: electrons (e−), photons (γ), muons (µ−), charged pions (π±) and protons (p). The network is a particularly useful tool for data analysis of particle interactions in LArTPC detectors, since the region of an interaction vertex often contains many particles. The MPID algorithm can take as input a LArTPC image with a fixed 512×512 pixel scale. A detailed description of the network design and training for MPID is given in Section 2. When used in MicroBooNE's deep learning based low-energy νe (LEE 1e-1p) search analysis, the MPID network is primarily applied to images that contain candidate reconstructed neutrino interaction vertices as well as all reconstructed topologically connected activity. MPID predictions are derived based on the full information of all energy depositions topologically connected to the vertex, particularly the first few centimeters of final-state particles' trajectories, which are critical for particle identification. In the νe search, the network is also applied to more inclusive images roughly cropped around the interaction vertex. This is a new feature compared with the single PID network, which takes as input only images containing filtered, reconstructed hits. Cropping around the interaction vertex allows re-evaluation of charge missing from the former topologically- connected image, but is nonetheless present near the vertex, such as photon showers from final- state π0s. This feature of the MPID network can help MicroBooNE suppress important photon backgrounds to a LEE search, as observed by MiniBooNE [23]. We demonstrate this feature's robustness against the presence of LArTPC activity such as cosmic ray tracks that are uncorrelated with signal features of interest. In this note, we demonstrate the expected MPID performance on simulated test images con- taining one muon and one proton (1µ-1p) and one electron and one proton (1e-1p) in Section 3. Both are signal interactions of a MicroBooNE deep learning based LEE analysis. We then demon- − strate the MPID's achieved level of µ and p identification in a filtered νµ charged current (CC) dataset and compare results between data and simulation in Section 4. We similarly demonstrate the MPID's e− and γ separation capability using a filtered sample of π0-containing events, and compare data and simulation results. In Section 5, we demonstrate the value of MPID for a LEE measurement in e− and γ separation, e− and µ− separation and p and cosmic ray separation expected from a set of simulated νe interaction images. 3 2 Multiple particle convolutional neural network 2.1 Network design The MPID network applies a typical CNN [24] structure for the task of multiple object classifi- cation, which is summarized in block diagram form in Fig. 1. Input images have a resolution of 512 × 512 (1.5 m×1.5 m) pixels, which generally matches the size of neutrino-induced activity in MicroBooNE. A series of ten convolutional layers are applied to the image for extracting high-level features. Figure 1: MPID network scheme. The output has five numbers. Each of the values is between 0 and 1, representing the probabilities of corresponding particles in the given LArTPC image. The first convolutional layer has a stride (shift unit of the convolution calculation) of two with the goal of reducing the LArTPC images' sparsity and increasing feature abundance at the beginning of the algorithm.